Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
3,784 questions
Incorrect "Safety+Security" policy violation error when parsing a PDF document with Azure Open AI
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(22.400000000000002, 52%, 12%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ED%3C/text%3E%3C/svg%3E)
Denise
0
Reputation points
We encountered an issue with Azure Open AI when attempting to parse and convert to markdown one of the pages of a PDF document that contained a family photo. In this case, Azure Open AI returned the following error message:
Error processing page: Error code: 400 - {'error': {'inner_error': {'code': 'ResponsibleAIPolicyViolation', 'content_filter_results': {'sexual': {'filtered': True, 'severity': 'high'}, 'violence': {'filtered': False, 'severity': 'safe'}, 'hate': {'filtered': False, 'severity': 'safe'}, 'self_harm': {'filtered': False, 'severity': 'safe'}}}, 'code': 'content_filter', 'message': "The response was filtered due to the prompt triggering Azure OpenAI's content management policy. Please modify your prompt and retry. To learn more about our content filtering policies please read our documentation: \r\nhttps://go.microsoft.com/fwlink/?linkid=2198766.", 'param': 'prompt', 'type': None}}
The PDF files our product parses can include photos, so we need to solve for this use case. Also, we were surprised by the filter as the photo didn't contain any sexual content. This looks like a false positive and it happened multiple times as we continued to try to parse the document.
We checked our content filters in the “Safety + Security” section and confirmed that the values for all parameters were already set to high for both input and output filters.
Can anybody help us with this?
{count} votes
-
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(6.4, 20%, 10%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EKS%3C/text%3E%3C/svg%3E)
kothapally Snigdha 1,715 Reputation points Microsoft External Staff2025-03-10T22:17:45.47+00:00 Hi Denise
It seems that you are encountering a "Responsible AI Policy Violation" error when trying to parse a PDF document with Azure OpenAI, specifically related to content filtering policies. This error indicates that the content management policy was triggered, resulting in the filtering of the response due to perceived high-severity issues in the "sexual" category, despite your assertion that the content does not contain any sexual elements.
- Ensure that the content being processed does not inadvertently contain elements that could be misinterpreted by the content filtering system. This includes examining the context in which the family photo is presented.
- Since the error message suggests modifying your prompt, consider rephrasing or simplifying the input to avoid triggering the content filter.
- If you believe this is a false positive, document the details of the incident, including the context of the content and the specific error messages received.
I hope these helps you! Thank you!
-
kothapally Snigdha 1,715 Reputation points Microsoft External Staff
2025-03-12T00:47:18.2+00:00 Hi Denise
Following up to see of the response was helpful.
-
Denise 0 Reputation points
2025-03-12T05:49:42.6433333+00:00 Thanks for your response. We believe this is a false positive. Where should we "document the details of the incident, including the context of the content and the specific error messages received"? We shared the error message on this post.
Sign in to comment
Sign in to answer